kmp-utils 1.0.2

Implementation of KMP algorithm and simple generalizations.

KMP Utilities

The KMP utils library provides a python binding to C++ for fast, linear time string processing. Currently available for linux only.

Installation

You can install the kmp_utils library with the following command:

pip install kmp_utils

Then import into your program as:

import kmp_utils

def main():
    s = "aabaaba"
    t = "ab"
    x = kmp_utils.split(s, t)
    print(x)
    pass

if __name__ == '__main__':
    main()

>>> ['a', 'a', 'a']

This library requires pybind11 and python >= 3

Examples

The kmp_utils library consists of 5 API methods.

find_all(s, t). Reading from left to right starting from the beginning of string s, find all disjoint occurrences of string t in s by returning the starting indices of any such occurrences. This returns an increasing list.

find_all("aaaaaa", "aa") = [0, 2, 4]
find_all("aabaaba", "ab") = [1, 4]
find_all("sdsdsd", "ab") = []

find_all_left(s, t). Reading from right to left starting from the end of string s, find all disjoint occurrences of string t in s by returning the starting indices of any such occurrences. This returns a decreasing list.

find_all_left("aaabbb", "aa") = [1]
find_all_left("aabaaba", "ab") = [4, 1]
find_all_left("sdsdsd", "ab") = []

get_next_right(s, i, t). Reading from left to right starting from index i in string s, find the next occurence of string t in s by returning the starting index. Returns -1 if t cannot be found.

get_next_right("aaaaaa", 5, "aa") = -1
get_next_right("aabaaba", 3, "ab") = 4
get_next_right("sdsdsd", 0, "ab") = -1

get_next_left(s, i, t). Reading from right to left starting from index i in string s, find the next occurence of string t in s by returning the starting index. Returns -1 if t cannot be found.

get_next_left("aaaaa", 1, "aa") = 0
get_next_left("aaaaa", 1, "aaa") = -1
get_next_left("ababaabb", 6, "ab") = 5

split(s, t). Split string s by t starting from the beginning of s.

split("aaaaa", "aa") = ['', '', 'a']
split("axbxcx", "x") = ['a', 'b', 'c']
split("ababaabb", "xs") = ['ababaabb']

Performance Testing

We compare a linear python iteration with the kmp_utils.find_all method with the following code.

import kmp_utils
import time
from typing import List

def python_kmp_find_all(text: str, pattern: str) -> List[int]:
    result = []
    prefixTable = computePrefixTable(pattern)
    index = KMPAlgorithm(text, pattern, 0, prefixTable)
    while index != -1:
        result.append(index)
        index = KMPAlgorithm(text, pattern, index + len(pattern), prefixTable)

    return result

def KMPAlgorithm(text: str, pattern: str, index: int, prefixTable: List[int]) -> int:
    n = len(text)
    m = len(pattern)
    if n-index < m or m == 0:
        return -1
    i = index
    j = 0
    while i < n:
        if text[i] == pattern[j]:
            i += 1
            j += 1
            if j == m:
                return i-m
            continue
        while j > 0 and pattern[j] != text[i]:
            j = prefixTable[j-1]
        if j == 0 and pattern[j] != text[i]:
            i += 1

    return -1

def computePrefixTable(pattern: str) -> List[int]:
    m = len(pattern)
    prefixTable = [0 for i in range(0,m)]
    j = 0
    for i in range(1,m):
        while j > 0 and pattern[j] != pattern[i]:
            j = prefixTable[j-1]
        if pattern[j] == pattern[i]:
            j += 1
        prefixTable[i] = j
    return prefixTable

def p1():
    n = 1000000
    s1 = 'a' * n
    s2 = 'a' * n
    p1 = 'a' * 10
    p2 = 'a' * 10

    t1 = time.time()
    x1 = kmp_utils.find_all(s1, p1)
    dt = time.time() - t1
    print(f'kmp_utils time: {dt} seconds')

    t1 = time.time()
    x2 = python_kmp_find_all(s2, p2)
    dt = time.time() - t1
    print(f'kmp algorithm in python time: {dt} seconds')

    assert(len(x1) == len(x2))

    for i in range(0, len(x1)):
        assert x1[i] == x2[i]


def main():
    p1()
    pass

if __name__ == '__main__':
    main()

>>> kmp_utils time: 0.009107112884521484 seconds
>>> kmp algorithm in python time: 0.40862512588500977 seconds

For coding interview preparation, please visit [algorithmspath.com] (https://algorithmspath.com).